One of the building services projects our WDR simulation consultants delivered was wind-driven rain for Canberra MRT station in Singapore. Wind-driven rain, as the name implies, is rain droplets propelled by wind blowing from various directions. WDR concerns public and private buildings because it can damage these buildings and reduce the occupants’ comfort and safety. A good example is above-ground train stations which in many cases were designed to rely on natural ventilation to increase passengers’ thermal comfort. When it rains, stations in which the design only addressed natural ventilation but not WDR mean rain droplets carried by the winds will ingress into the station. An approach to minimise if not prevent the ingress of WDR is to address it using computational fluid dynamics (CFD) modelling & simulation.
The objective of performing WDR simulations is to determine the effectiveness of a building’s weather protection features whilst still maintaining the effectiveness of its natural ventilation capability. Performing WDR simulation using CFD will address these questions: Is the weather protection effective against WDR? If rain droplets can penetrate the interior spaces, where are the exact locations, what is the extent, distance and quantity? Which façades are affected the most? Even if the weather protection is effective against WDR does it compromise the natural ventilation of the station? How effective are the louvres in allowing sufficient air to enter the interior space whilst keeping the rain droplets out? Thus, the objective of WDR simulation is to minimise the ingress of WDR into the interior space of the station in order to increase the comfort, enhance the safety of the passengers and prevent rainwater-induced damages to the station.
To simulate WDR, typically four rain droplet sizes, each with its own terminal velocity and drag coefficient, will be modelled and simulated. The reason for simulating rain droplets of various sizes is because each of these rain droplets responds differently to wind. For example, smaller rain droplets tend to be carried by the wind and follow its direction whereas larger heavier droplets, due to their mass, tend to travel based on their own trajectories, which are mostly straight lines and are less affected by the mean flow of wind. Prevailing wind in eight directions distributed radially at a 45 degrees angle, obtained from Australian Bureau of Meteorology, will be modelled and simulated. An approaching train will also be factored into the numerical model since it can pull the rain droplets into the interior space of the train station. It is well established that, when simulating natural ventilation, it is important to include neighbouring buildings because these buildings can block the wind and alter the wind direction. However, in WDR, consideration should be made to model the building of interest by itself without neighbouring buildings to simulate the highest wind velocity if the existing neighbouring buildings are pulled down in the future and they should also be modelled and simulated with neighbouring buildings as well because these buildings can alter the direction of wind and thus the rain droplets trajectories. From here, the results from the worst of the two scenarios can be employed to fine tune the station’s design.
Modelling & simulation of wind-driven rain via computational fluid dynamics can help the design engineer to achieve a balance between the ingress of WDR into the station and passengers’ experience from using the station. Most importantly, if conducted as part of the building envelope assessment during the design phase, it can prevent costly rework of the station post-construction.
Buildings that have been designed for natural ventilation will save on energy bills whilst keeping their occupants thermally comfortable. For a building to be effectively natural ventilated, it has to be designed to be one in the first place. CFD modelling and simulation is a powerful technology employed to provide guidance to architects and engineers with respect to the design and construction of highly effective naturally ventilated buildings.
When modelling ambient wind flowing around a building, our fluid domain will include the development of interest and the far field boundary which will be located far enough from the building model to avoid artificial acceleration of the flow.Our standard practice of simulating natural ventilation is to model and simulate the building of interest including the neighbouring buildings, structures and/or natural landscapes because they have the potential to partially block the wind and thus alter the flow field resulting in increased pressure drop, tunnelling effect (localised high velocity), decreasing or increasing the amount of wind entering the building of interest. Hence, the wind profile depends on the size and arrangement of neighbouring buildings.This inclusion of the neighbouring buildings, structures and/or natural landscapes is performed to ensure a more realistic scenario is simulated. To be specific, the extent of the surrounding buildings to be explicitly modelled shall be within the proximity of minimum 3 times the length of the longest distance measured across the boundary of the development, or within 500m distance from the edge of development of interest, whichever that is smaller. In the event that the building and surrounding development are located within hilly terrain, the topography information shall also be included in the simulation models. The domain height shall be extended, approximately 6 times the height of the tallest building within the defined vicinity. A project we delivered for the Singapore Housing Development Board (HDB) was described and illustrated below.
To determine the most effective building design and layout for natural ventilation, we performed macro and micro CFD modelling and simulation for the site development. The macro simulation involved large scale ventilation simulation to determine the wind flow condition around the proposed building development and neighbouring buildings. On the other hand, micro simulation involved simulating wind flow condition of all living spaces inside the units. During the simulation, all windows and doors were assumed to be fully opened as designed except for the main door which was assumed to be closed at all time.
Data Centre Room Layout Optimisation
The objectives of performing CFD modelling and simulation on Reserve Bank of Australia data centre located in Sydney were to: (1) determine the effectiveness of the thermal cooling system by analysing the temperature, air flow and pressure, (2) optimise the data centre room layout design, (3) predict the uniformity of air flow coming out from the tile diffusers and (4) predict if a CRAC unit fails, how much time was available for the operators to react.
Our accurate CFD simulation results answered 4 important questions:
Cooling system – we determined the presence of temperature hotspot, air flow dead zones and cooled air under or oversupply. The effectiveness of a data centre thermal management system was determined by its ability to maintain a consistent homogeneous room temperature profile and prevented the room temperature from exceeding certain temperature.
Optimisation of layout – answered ‘what if’ scenarios such as the relocation of tiles, server racks, server rack maximum height and CRACs returns. It was always so much financially cheaper to investigate the impact of changing a variable in the ‘virtual model’ than to do so in the real physical world. For example, the impact of adding server racks on the overall cooling capacity was simulated first to ensure sufficient cooling would still be available and no additional hot spots were being introduced. After investigating the relevant variables, the data centre was modified with confidence and cost effectively.
Uniformity of air flow – predicted the air flowrate and temperature coming out from each tile based on sub-plenum design and features such as piping, cable trays and support columns. Knowing the flowrate coming out from each tile enabled the tile to be positioned in appropriate locations which in turn prevented the oversupply or undersupply or cooled air to the server racks.
Response time – if a given CRAC unit fails, predicted the duration available for the operators to react. Knowing the response time available meant a contingency plan that factored in this response time could be developed to prevent the room from overheating and damaging the server equipment.
Pollutant (plume) Emission
The objectives of performing CFD modelling and simulation of plume emission in Lee Kong Chian School of Medicine located in Singapore were to determine the effectiveness of the pollutant discharge equipment and to optimise the system so that the emission would not recirculate to fresh air intake units of the building and affect neighbouring buildings. In addition, there were other exposure limits set out by OSHA PEL, NIOSH REL, ACGIH which defined the permissible exposure limits for a given exposure period, as well as other critical concentrations such as odour threshold limit for the case of ammonia smell nuisance. This LKC School of Medicine building was surrounded by other high rise apartments, thus, it was particularly important to ensure that the designed pollutant discharge was able to meet the required emission & exposure limits.
Recirculation into air intake units: in building and construction industry, it was common to locate both fresh air intake units and plume discharge equipment in close proximity such as on a building roof. Hence, there was possibility that discharged pollutant was drawn backed into the building through the air intake units. This could be the result of insufficient plume height and locations of the equipment.
Impact on neighbouring building: due environmental wind conditions, pollutant dispersion could be carried over to neighbouring buildings, which might have air intake units on their roofs or occupants with rooms exposed to the plume. If the plume height was not high enough to overcome neighbouring buildings, these plumes would have negative consequences on these buildings. This was especially so for taller neighbouring buildings or neighbouring building on a higher topography.
Simulation scenarios: multiple scenarios were considered to study the impact of different variables. These variables included wind direction, wind speed, variable performance level of discharge equipment and ambient temperature conditions. An example of such a project we did was the Singapore LKC school of medicine.
As building services consultants within the building & environment industry, we provide performance-based building services design consulting in the following areas:
- Building acoustics analysis
- Building performance
- Cooling tower exhaust ventilation
- Data centre room layout optimisation
- Generator exhaust extent of recirculation and HVAC ducts optimisation
- Natural and mechanical ventilation simulation studies
- Optimisation of solar chimney design
- Performance-based fire safety design
- Pollutant (plume or particles) emission
- Wind-driven rain
Buildings such as an airport consumes much energy and contribute to greenhouse gas emissions. The quests are thus to reduce the carbon footprints of buildings and their energy usage. Since buildings are relatively expensive to build, their performance must be known and ensured prior to equipment procurement and the commencement of construction. Our building services consultants performed building CFD simulations on the entire Jakarta Terminal 3 Airport to analyse impact of changes, answered 'what if' scenarios, provided comprehensive data which were difficult to obtain from experimental tests and allowed alternatives to be studied before optimum designs were confirmed. In the end, a beautiful, world-class airport that offers great comfort to the passengers was built and commissioned.
As building services consultants with expertise in building performance analysis, our in-depth building cfd simulation services have assisted our clients to achieve optimum building performance in the areas of thermal comfort analysis (velocity, temperature, RH), design of natural ventilation, solar and daylighting analysis, embodied energy, energy usage, architectural design, room layout, HVAC sizing for occupants comfort and optimising the locations of diffusers and returns.
Cooling Tower Exhaust Ventilation
The objectives of performing CFD modelling and simulation of cooling tower exhaust in OUE building Singapore were to determine the extent of exhaust recirculation into the cooling towers and to establish the temperature profile around the vicinity of cooling towers. Our CFD simulation results addressed the following issues:
Recirculation into air intake units: in building and construction industry, due to the space constraint on the building roof, it was common to locate the cooling towers in close proximity to one another. Moreover, trellis were installed above these cooling towers for aesthetic reason. Because of such a scenario, there was a possibility that the exhaust from a cooling tower recirculated into its own air inlet or inlets of other cooling towers. This might eventually affect the performance of the cooling towers. Thus, our CFD simulations predicted whether recirculation occurred based on the proposed arrangement and locations of the cooling towers.
Temperature profile: a temperature profile around the cooling towers was also established to determine whether elevated temperature affected the cooling towers performance or whether they were acceptable to the efficient operations of the cooling towers.
Impact by neighbouring buildings: due environmental wind conditions, the performance of a group of cooling towers might be affected by neighbouring buildings. This was especially so for shorter buildings being surrounded by taller neighbouring buildings which might act as obstacles to wind flow.
Simulation scenarios: multiple scenarios were considered to study the impact of different variables. These variables included wind direction, wind speed, variable performance level of discharge equipment, ambient temperature conditions and height of trellis. An example of such a project we did was the Singapore OUE building.
Performance-Based Fire Safety Design
The objectives of performing CFD modelling and simulation of performance-based fire safety design on the AEON Shopping Mall based in Jakarta were to characterise the behaviour of fire, smoke and ultimately determine the robustness of the proposed fire safety design. Our CFD modelling and simulation results assisted the building designers to enhance the performance-based fire safety design and addressed the following issues:
Enclosure: defined the enclosure being studied in the CFD modelling. The enclosure details included the identity of the enclosures that belonged in the fire model analysis, the physical dimensions of the enclosures included in the fire model, and the boundary conditions of each enclosure.
Fire Characteristics: the source fire was the critical input for the fire scenario and was often described as the ignition source. The source fire was typically characterised by a heat release rate although other important aspects included the physical dimensions of the burning object, its composition, and its behaviour when burning. The heat release rate could have been specified as a continuous function of time or it might be an array of heat release rate and time data.
Computational Domain: the computational domain was as close as reasonably practicable to the actual enclosure. Where inlet and/or exhaust vents were located at the domain boundaries, we included an additional 5m buffer outwards to account for the aerodynamics of the vents. Where wind effects were being modelled, the domain was extended correspondingly to take this into consideration. Temperature: in general, ambient temperature in air-conditioned spaces was taken to be 24 deg Celcius while that of non air-conditioned spaces, 32 deg Celcius. Smoke Management Systems: relevant smoke management system i.e. engineered smoke control system, smoke purging system or smoke vents was included in the CFD modelling.
Report: the report included output quantities for visibility, temperature and velocity. Additional output quantities e.g. carbon monoxide, heat flux, FED could also be included in the report but this depends on the scope of the performance-based design. For slice parameter, slices in all three planes (x, y and z-planes) were included. A cut along the centreline of the fire origin and slices at critical areas in the model were included. More slices could be included if a spill plume forms part of the analysis. At least 2 slices of z-plane at 1.7m and 2.5m above the finished floor level were included. Additional slices at critical areas in the model could be included if required. An example of such a project we delivered was the Jakarta AEON shopping mall, located in Indonesia.
AUSTRALIA SINGAPORE ASIA PACIFIC REGION