Level of abstraction and of detail ? Simulation has to decide what is important and what is not.
Depending on the problem, focus on the different aspects of building "energy/mass" balance.
In this lecture the student will learn about specific modes of energy transfer in the building straucture and the application of simulation and modelling.
You can access more information under a variety of topics:
Extract from text book:
"Energy Simulation in Building Design"
J A Clarke, 1985
Figure 1.1 shows the various flowpaths commonly encountered within and outwith buildings and which interact, in a dynamic manner, to dictate comfort levels and energy demands. Underlying these flowpaths is the concept of energy, mass and momentum balance which requires, in turn, a knowledge of the fundamental processes of conduction, convection and radiation exchange. Many excellent texts exist which cover the fundamentals of heat transfer [for example, Kreith 1973, Ozisik 1977, Incropera et al 1981] and no attempt is made here to commence this book with an elementary treatment of the subject. Instead specific mathematical models are introduced for each flowpath and, as the book develops, these are combined to form a single, unified mathematical structure which is the statement of whole system balance. It is necessary therefore to commence with an explanation of the flowpaths encountered in building systems and so candidates for inclusion in the simulation model.
Figure 1.1 Energy Flowpaths
This is the process by which a fluctuation of heat flux at one boundary of a solid material finds its way to another boundary, being diminished in magnitude due to material storage, and shifted in time. Within the building fabric, transient conduction is a function of the temperature and energy excitations at exposed surfaces, the temperature-dependent (and therefore time-dependent) thermophysical properties of the individual homogeneous materials, and their relative position. For modelling and simulation purposes it is usual to declare external climatic excitations as known time-series data with the objective to determine internal transient energy flow and hence the dynamic variation of heat flux at internal surfaces.
The thermophysical properties of interest include conductivity (W/m°C), density (kg/m3) and specific heat capacity (J/kg°C) as well as physical dimensions. These properties are themselves time-dependent because of intra-material temperature and/or moisture fluctuations. However, in many applications, such dependancies are ignored and properties are assumed to be invariant in the time dimension. Materials with high thermal diffusivity values (conductivity divided by the product of density and specific heat; units m/s) transmit boundary heat flux fluctuations more rapidly than do materials with correspondingly low values. Appendix A lists the basic thermophysical properties of a range of building materials. Listed properties include conductivity, density, specific heat, surface emissivity and solar absorptivity.
The relative position of the different materials within composite constructions can also greatly influence transient behaviour. Traditionally designers have relied on the simple steady state U-value concept to assess the heat loss characteristics of the building fabric. In addition to ignoring the dynamic aspects of fabric behaviour, this approach does not preserve the spatial integrity of a multi-material construction since in reality different constructions can perform differently although each may have the same U-value.
As an example, if insulation is located at the innermost position of a wall then any shortwave solar radiation penetrating windows and striking that internal surface cannot be readily stored in the construction since the insulation will act as a barrier. Instead, the solar energy will cause a surface temperature rise which, in turn, will increase the rate of energy release to the adjacent air by the process of natural convection. A space experiencing high solar energy penetration is therefore likely to overheat if cooling is not introduced in some way. Conversely, if the insulation is relocated externally with capacity elements exposed to the inside then internal surface shortwave gain can access capacity to be stored. By proper design this stored energy can later be harnessed (passively rather than by mechanical means) to minimise heating requirements rather than cause overheating. On the other hand internal capacity may give rise to increased peak plant demand due to the initial rush of energy to capacity at plant start-up in an intermittent scheme. With continuous operation, capacity can help to minimise the peaks and maximise the troughs of plant demand and so promote good load levelling. This in turn will give a stable environment and encourage efficient plant operation by allowing plant to operate consistently at or near full load. The risk of interstitial condensation is however greater in the case of internally located insulation since a substantial portion of the construction may fall below the dew point temperature of moist air permeating through the construction in the absence of an effective vapour barrier.
In summary, transient conduction will affect:
Unfortunately there is no simple design paradigm which can be used to select an optimum construction. The table below, for example, shows the effect on cooling energy requirements, peak plant demand and load levelling for a number of combinations of construction and plant operation. If these results are accepted then it is clear that there is a need to utilise dynamic models to determine the performance of alternative constructions when each are combined with the many other combinatorial design features: whilst a U-value can be used as a simple selection index, it has little if no role as a sophisticated indicator of construction energy performance.
| Construction# | Energy (kWh) | Load Levelling (LL) (Qmax-Qmin; kW) | Note |
| Continuous Heating System Operation | |||
| 40I/150C/40I | 410 | 22 | LL best when more insulation to outside of capacity |
| 5I/150C/75I | 396 | 18 | |
| 75I/150C/5I | 413 | 25 | |
| 5C/80I/75C | 398 | 17 | LL best when more capacity to outside of insulation. |
| 25C/80I/125C | 411 | 20 | |
| 125C/80I/25C | 391 | 16 | |
| Intermittent Heating System Operation | |||
| 40I/150C/75I | 268 | 37 | LL best when insulation split equally either side of capacity. |
| 5I/150C/75I | 305 | 39 | |
| 75I/150C/5I | 275 | 43 | |
| 75C/80I/75C | 330 | 44 | LL best when more capacity to inside of insulation. |
| 25C/80I/125C | 336 | 53 | |
| 125C/80I/25C | 328 | 42 | |
| # Each Cconstruction comprises 3 homogeneous layers each of which can be insulation (I) or capacity (C). The specifications are all inside to outside and all constructions have the same U-value. The dimensions are in mm. | |||
This is the process by which the heat flux, eminating at some opaque or transparent surface, is transmitted to an adjacent fluid layer. For building surfaces it is usual to differentiate between external and internal exposures. In the former case convection is usually considered as forced whereas, with internal surfaces, natural and/or forced air movement can be assumed depending on the available information on mechanical equipment and the convective field to result.
It is normal practice in simulation modelling to make use of dimensioned convection coefficients (W/m°C) which represent some average value for a particular finite surface area but can change with time. In the fluid flow literature it is common to express fluid to surface heat transfer in dimensionless terms but, in recent years, some workers [Alamdari et al 1982] have recognised the needs of the modelling community and produced correlation expressions to allow estimation of time-dependent but surface-averaged convection coefficients of much use in simulation applications.
Forced convection is a function of the prevailing fluid flow vector. Typically, for external building surfaces, wind speed and direction data is available for some reference height and simple techniques exist to estimate non-reference height values in terms of characteristic vertical velocity profiles. Forced convection estimation for internal surfaces is more problematic requiring knowledge of the distribution and operation of air handling equipment and heat emitters and the nature of the boundary layer at each surface position (laminar, turbulent or transitional).
Natural convection is an easier problem to study and many formulations have emerged which express coefficients as simple functions of the surface-to-fluid temperature difference; surface aspect, roughness and dimension; and direction of heat flow.
In many calculation methods surface heat transfer coefficients are treated as combinations of convection and longwave radiation although the values used are often dubious. In reality the two processes are related by the fact that they both conspire to raise or lower surface temperatures and so influence each other.
Inter-surface longwave radiation is a function of the prevailing surface temperatures; the emissivity of each surface; the extent to which the surface pair are in visual contact, often referred to as the view factor; and the nature of the surface reflection, specular or diffuse. The flowpath will tend to establish surface temperature equilibrium by cooling hot, and heating cold, surfaces. It is most important under conditions of asymmetric heating often found within passive solar applications in which an attempt is made to capture shortwave solar energy at some elected surface.
The mathematical representation of the flowpath is non-linear in the temperature term and this will introduce complications in simulation modelling where the condition of linearity is desirable.
The exchange of energy by longwave radiation between external (opaque and transparent) surfaces and the sky vault, surrounding buildings and ground can result in a substantial lowering of surface temperatures especially under clear sky conditions at night. This process alone can result in sub-zero surface temperatures especially in exposed roofs and can become critical in cases of low insulation level. Conversely the flowpath can result in a net gain of energy although under most conditions this would be negligible.
From a modelling viewpoint an adequate treatment of this flowpath will require the ability to estimate the effective sky temperature as a function of prevailing cloud cover and type; the temperature of surrounding buildings; the temperature of the ground as a function of terrain conditions, air temperature and incident shortwave energy; and the relevant view factor information which visually couples the surface with the three portions of its scene.
In most buildings the gain of energy from shortwave penetration constitutes a significant portion of the total loading and therefore the method of treatment of shortwave flowpaths can largely determine the accuracy of the overall predictions.
Some portion of the shortwave energy (arriving directly from the sun or diffusely after atmospheric scatter and terrain reflections) impinging on an exposed surface may - depending on subsequent temperature and energy variations affecting transient conduction - eventually find its way through the structure where it will contribute to the inside surface heat flux at some time later. It is not uncommon for exposed surfaces to be as much as 10-15 °C above ambient temperatures. Many existing techniques utilize the concept of a 'sol-air' temperature which represents some suitably elevated ambient temperature for use as the index in subsequent wall conduction calculations. This is clearly inadequate on two points:
In the case of completely transparent structures, the shortwave energy impinging on the outermost surface is partially reflected and partially transmitted. Within the layers and substrates of the system many further reflections take place and some portion of the energy is absorbed within the material to raise its temperature. This temperature rise will augment the normal transient conduction processes and by this mechanism help to establish extreme innerside and outerside surface temperatures which then, in turn, drive the surface convective and longwave radiative flowpaths. Thus, in effect, absorbed shortwave radiation penetrates the building via convection and longwave radiation.
The component of the incident beam which is transmitted will eventually (with no perceptible time lag) strike some internal exposed surface or surfaces where it behaves as did the external surface impingement: opaque surface absorption and reflection, transparent surface reflection and transmission (to outside or another zone), and behind-the-surface transient conduction where it is stored and lagged.
Accurate solar modelling therefore requires a number of algorithmic methods for the prediction of surface position relative to the solar beam as well as exposed surface shading and the moving pattern of insolation of internal and external surfaces. Surface/solar position is a function of site latitude and longitude, time of day and year, and surface geometry. Accurate shading/insolation estimation requires the existence of ray tracing techniques of the kind derived in chapter 5.
The thermophysical properties of interest include shortwave absorptivity for opaque elements and absorptivity, transmissivity and reflectivity for transparent elements. The magnitude of these properties is dependent on the angle of incidence of the shortwave beam and on its spectral conposition. With regard to the latter, it is common practice to accept properties which are averaged for the spectral portion under consideration.
These processes control the magnitude and point of application of solar energy and so dictate the overall accuracy of any solar processing. Both time-series are usually expressed as proportions of one or as percentages and will require sophisticated point projection or hidden line/surface techniques for their estimation as well as access to a data structure which contains obstruction features.
It is usual to assume that facade shading caused by remote obstructions (such as buildings, trees etc.) will reduce the magnitude of direct insolation leaving the diffuse beam undiminished. Conversely, shading caused by facade obstructions (such as overhangs, window recesses etc.) should also be applied to the diffuse beam since the effective solid angle of the external scene, as subtended at the surface in question, may be reduced.
At any point in time the shortwave radiation directly penetrating an exposed window will be associated with one or more internal surfaces, depending on the prevailing solar angle relative to the window and internal building geometries. Thus the receiving surface(s) may be an opaque surface, a window in another wall (connecting the zone to another zone or to ambient conditions), items of furniture, and so on, depending on the established data structure. While it is true that disregarding the proportioning of window transmitted shortwave energy between the associated receiving planes can have a significant effect on both the quantitative and qualitative aspects of thermal predictions, the smearing of the portion received by one surface over its entirety will have little quantitative effect if the surface can be regarded as uniform in the sense that it is of the same composition with no spatially-dependent boundary conditions [Robinson 1979]. Likewise there is no appreciable qualitative effect for the case of uni-directional conduction heat flow representations.
In building simulation modelling two fluid flowpaths predominate: infiltration and zone-coupled air flow. And these flowpaths give rise to advective (fluid to fluid) heat exchanges. Both are vector quantities in that only air flow into a region is considered to cause thermal loading, any air loss merely being the driving force for a corresponding replacement to maintain a mass balance.
Infiltration is the name given to the leakage of air from outside and can be considered as being comprised of two components: the unavoidable movement of air through distributed leakage paths such as the small cracks around windows and doors, through the fabric itself and at material junctions; and the ingress of air through intentional opennings (windows, vents etc.) often referred to as natural ventilation.
Zone coupling, like infiltration, is caused by pressure variations and by buoyancy forces caused in turn by density variations due to the temperature difference between the coupled volumes of air.
Thus random occurrences such as window and door opening and changes in the prevailing wind conditions or the intermittent use of mechanical ventilation will have some effect on infiltration, natural ventilation and zone-coupled air flow. Although the effect (of these occurrences) on air movement is difficult to determine, models of varying complexity can nevertheless be constructed. Such models will span the spectrum from simple whole building predictors based, say, on linear regression methods to complex simulation systems involving a numerical solution of some governing partial differential flow equation.
At a level appropriate to building energy modelling, air movement is often represented by a simplified nodal network in which nodes represent volumes and nodal connections represent the distributed leakage connecting the volumes and through which air movement may occur. Numerical techniques can then be applied to this network to establish the mass balance corresponding to any particular nodal pressure field.
In many buildings the effects of heat gains from lighting installations, occupants and miscellaneous equipment can be considerable and it is therefore important to process these heat sources in a realistic manner. This will entail the separate processing of the radiant and convective components and the ability to work with time-dependent profiles to allow any casual source to change on an hourly, daily, weekly and/or seasonal basis. It is usual to assume that the convective component is experienced instantaneously as an air load but that the radiant portion, behaving in a similar manner to shortwave radiation penetrating the building envelope, is apportioned among internal opaque and transparent surfaces according to some angular distribution strategy and so has a relationship with system capacity and will be lagged.
The problem of predicting the energy consumption of a building is usually divided into two distinct stages. As shown in figure 1.2, the first stage is concerned to predict the energy requirements to satisfy the demands of the building activity. This is found by in some way modifying the various instantaneous heat gains and losses by the ever present thermal storage and lag effects. In the second stage these energy requirements are modified by the operating characteristics of the selected plant to give the energy actually consumed. Thus the first stage is concerned with the design of the building to reduce the energy requirements (and so, perhaps, consumption) whilst the second stage is concerned with the design of the installed plant to best match these requirements and minimise consumption. This is the strategy underlying many of the simulation systems presently available.
It is possible however to develop systems which can model the combined building/plant configuration in a simultaneous manner. This is demonstrated in chapter 6 where selected plant systems are combined with the building model developed in chapter 3. Irrespective of the level of plant modelling, the building model will be required to handle control statements which superimpose complex time and thermostatic constraints on the availability of plant capacity to the building system.
Also, it has been the general practice to control simulations on the basis of simple tests applied to zone air temperatures or some equivalent single index and to declare plant in a conceptual manner such as convective (air volume interaction) or mixed (air and surface interaction). It is now possible (and often desirable) to operate in terms of actual controller sensed temperature determined on the basis of air movement and radiant effects assessed from subtended solid angle considerations. It is also possible to allow plant to interact with building regions in a manner which more closely adheres to the reality.
Fluctuations in moisture level will obviously affect cooling loads in systems which permit humidity control. It can also affect the thermophysical properties so often assumed constant for building modelling purposes; an assumption difficult to justify for plant modelling and in certain building applications such as passive solar designs where temperature variations within some components may be considerable.
In recent years many designers have come to favour the use of the so-called passive solar features. These act to capture and process solar radiation passively and without recourse to mechanical systems. Consider figure 1.3 which summarises the main passive solar elements. In each case certain factors can be identified which, in particular, will impose technical complecity on any modelling exercise. These are:
Advanced modelling systems seek to include each of the energy flowpaths while respecting the inevitable interactions and underlying complexities.
It is impossible to establish, a priori, the optimum level of model accuracy and flexibility in the field of energy systems appraisal. Indeed the trade-off between accuracy and flexibility is itself a dynamic concept which will vary according to the modelling and design objectives. Nevertheless it is important to differentiate between simplified models and comprehensive models which are capable of simple model emulation.
In the former case a number of simplifying assumptions are applied to the underlying thermal network and/or solution scheme so that some flowpaths are omitted entirely or approximated. The model to result is then valid only when applied to problems which embody the same or near-same simplifications.
In the latter case a comprehensive model is designed to operate on input data ranging from 'simplified to detailed' depending on the application in hand. This is achieved by incorporating 'dynamic defaults' to allow the inclusion of any flowpath not explicitly addressed in the input data set.
The latter model is obviously more flexible with the accuracy level changing as a function of the quality of the design information supplied. The following example is included to demonstrate the consequences of the two approaches.
Following on the introduction of the 1978 Building Regulations for England and Wales, a study was undertaken [ABACUS et al 1979] to examine the consequences of compliance or non-compliance with the 'deemed-to-satisfy' provisions. Regulation FF3 addressed the 'Conservation of fuel and power' and stated:
"A building or part of a building to which this part applies shall be so designed and constructed that the enclosing structure provides adequate resistance to the passage of heat the loss of which from the building or part would entail the consumption of fuel or power to enable temperature conditions normal for the proposed use of the building or part to be maintained."
Two approaches to complying with FF3 are set out in the provisions of FF4:
The study team felt that it was important to draw a distinction between prescriptive and performance requirements. The deemed-to-satisfy provisions of FF4, by focussing on heat loss rate per square metre of fabric, prescribed allowable construction in large measure, thus precluding innovatory facade treatment. More worryingly, it is entirely possible to satisfy the provisions with a design which, in terms of geometry, thermal mass, insulation, orientation, plant control, etc, may be profligate in energy consumption. Had the provisions dealt directly with performance - maximum annual energy consumption based on typical occupancy and operational statistics - the onus would be on the designer to present a design solution, however innovatory, together with appropriate predictive evidence of its energy behaviour. The issue then, if such a performance concept is accepted, is one of modelling accuracy and flexibility.
As part of the study a hypothetical but entirely typical test case, a multi storey hotel complex, was subjected to rigorous simulation to determine annual energy requirements against alternative glazing scenarios. Figure 1.3 gives the results (the solid line) and demonstrates that areas of glazing greater than the deemed-to-satisfy limit (25% single glazing in this case) can offer a significant saving in cost-in-use terms (point C-4 compared with point R-1, the latter obtained by subjecting the regulation limit scheme to the same simulation). In other words, the provisions by their apparent exclusion of building geometry, orientation, thermal inertia, shading, climatic variability, etc, may limit a designer to a solution which is somewhat removed from the optimum in terms of energy consumption and comfort performance.
Also shown on figure 1.4 (the chain lines) are the results obtained from two of the more commonly employed techniques: the RIBA calculator method and manual calculation methods taken from the CIBS Guide. These results raise an additional point: if in certain cases effective energy management can only be achieved by going beyond the constraints of the regulations then the designer cannot rely on the available simplified methods to provide the necessary evidence of performance since the results so obtained may be inaccurate and therefore misleading. And, of course, this problem is further compounded if additional technical complexity is introduced such as with passive solar elements or advanced control systems.
As a general strategy it would seem reasonable to aim for a high level of accuracy combined with a model structure which is capable of adapting to the information available at any design stage. It is likely that a truely 'simple' model, as perceived by a user, will be internally comprehensive in its treatment of the energy flowpaths, relying on the proper design of the software/hardware combination for its operational flexibility. This is the philosophy underlying the modelling approach developed in this book. The contention throughout is that accurate and flexible appraisal models can only be achieved by an approach which:
These represent the three pole axiom of conservation of energy, conservation of integrity and conservation of flexibility, which is the essential target of modelling systems.