The first of the 3 Laws of Thermodynamics, known also as the Law of Conservation of Energy, states that energy can neither be created nor destroyed. It only changes form (apart from at the Quantum level, but let’s not get into that!).
The way in which energy in a building changes form and moves is of great interest to those designers who use this information to best plan construction projects. The thing is, these are really complex system. With the capacity of computers to perform a multitude of calculations (a task in this context practically impossible by humans), digital simulation software can help designers to understand vastly complex relationships of cause and effect between building materials, humans and the environment.
Environmental Design in our built environment
Environmental design has existed in buildings since the first cave dwellers used geothermal heat from the earth to keep warm, to the use of natural ventilation and the storage of the Suns energy with thermal mass walls in vernacular architecture.
An example of environmental design is the positioning of buildings, and particularly their windows, to orrient towards the Sun, occupants can benefit from free energy in the form of solar gain.
In order to understand the behaviour of buildings, environmentally speaking, designers perform digital simulations to study heat gain, air flow directions, thermal performance of the building materials and the effects of occupants and energy producing equipment on the internal environment.
Testing multiple design options
One way that building energy simulations may be used in an optimised design environment is to compare various design alternatives. For the purposes of the Proleek project, three design options will be compared for the restoration of the cottage, based on suitable designs for the building type, traditional solid stone wall. These are categorised as:
1) Minimal intervention
2) Medium intervention, and
3) Considerable intervention
Following the preparation of the design models in Autodesk Revit 2016, using the ‘design options’ function (allowing for multiple variations on arrangement) the Architectural Building Information Model will be transferred to simulation software through the open source gbXML file format which translates spaces and surfaces into a simulateable form.
The spatial and thermal properties of the designs will be used directly in both EDSL TAS Simulator and Solar Computer. This step will be dealt with in detail in a later post, but suffice to say that for the research project, the aim is to automate as much as possible in this workflow, allowing for many itterations leading to optimised design.
The image below shows the various ways that heat and energy are transferred in buildings. The calculation of the effect of all of these processes combined is known as a thermodynamic simulation.
For the thesis project I will be using EDSL TAS Simulator to learn the heating and cooling loads throughout the year. Solar Computer will then be used to calculate the worst case scenarios on the hottest and coldest days of the year, for the purpose of sizing the capacity of the energy producing systems (mechanical or other… ie. fireplace?) required to sustain thermal comfort in the building.
Initial results for design option 1
The results are in!!! An initial test run has been completed for the first option of the Proleek cottage which sees the building resorted to its original form with no additional insulation. The simulation in EDSL TAS Simulator calculates the thermodynamic movement for every hour over an entire year based on statistical weather data for the region. The colours represent the energy within the space. We can see the Sun arc moving accross the sky as the year progresses and falling again towards winter. This shows us the heat naturally captured by the building from the environment.
On a final note, there are 8760 hours in a year…
I have learned this due to the following critisism. No doubt that the EDSL TAS Simulation software is very powerful and the output is in the form of rather attractive 3D environment on top of the usual graphs, charts and spreadsheets. Having said this, the program does not have a function to record or ‘play’ through the simulation. In order to create the above video I had to click very quickly 8760 times (taking a solid 26 minutes!) through each hour of the day while recording the screen (this was later sped up x8, ’cause who’s got 26 minutes?) It would be a welcome adition to the software!
Last week I was in Venice for the Architecture Biennale, a Bi-annual architectural event, showcasing the achievements of individual countries in the fields of architecture, construction technology and materiality.
While meandering through the canals of Venice I pondered the challenge of how to capture all of the cities architectural information for future planning and preventatively maintaining the slowly sinking city, a Unesco world heritage site.
It seemed to me that we are a long way away from a truly integrated way to capture reality. How can we document and manage all of the nooks and crannies of this ornate medieval city on stilts in the Adriatic lagoon?
I turned back onto the canal where I was staying in the Cannaregio district, and happened upon a Faro laser scanner, spinning away. It unfolded that a team from the Venice School of Architecture were conducting a research project on the reality capture of Europe’s oldest ghettos, formerly occupied by Jews, compelled to live there under the Venetian Republic.
Students painting and sketching on the canal at Cannaregio
It was interesting to notice the contrast of artists painting, the oldest form of capturing building geometry, and the laser scanner,picking up half a million spatial points every second!
Although our ability to capture precision building geometry has become almost fully automatic in recent years with the advent of laser scanning, it is what we do with this gathered data afterwards, which can have significant effects on the quality of restoration projects. There is a large gap in the process of how we extract this spatial data into useable digital building models for construction planning. Current manual processes are time consuming (expensive), require specialist skills and are prone to human error.
As part of the research project, a comparison of manual and automatic extraction of semantic 3D geometry from the point cloud, was undertaken. Below is the software workflow for each of the compared processes.
For the manual process, point cloud data was brought into Autodesk Revit as an external reference (or X-ref for my technical readers!) By viewing the point cloud of the cottage from different directions and section cuts, 3D geometry was modelled on top to a reasonable accuracy for the building type. During the process, generalisations about plumb-ness and uniformity were made. In other words, the building was straightened up somewhat, a limitation in the software at present if the geometry is to be useful afterwards.
In the automatic extraction process, the leading point cloud extraction tool was tested (Kim et al, 2015). Edgewise 3D is a software application which uses advanced algorithms to recognise and classify building components in a point cloud, and extract them to BIM authoring software as an ‘as built’ model. Edgewise 3D is primarily used in the mechanical engineering discipline (MEP) which benefits most from the recognition of many parts. It is possible to extract entire scanned plant-rooms in a matter of minutes. This is a game changer when we consider that the traditional method can take literally months, with huge windows of opportunity for human error.
Edgewise picked up on levels and surfaces in the point cloud and allows the user to ‘find’ walls, windows and ground terrain, and export this data to BIM authoring software. For comparison in this case, Autodesk Revit was again used. With the Edgewise plugin for Revit, the user is provided native Revit families, system families and height locations.
In all, the Edgewise automated process was able to go as far as giving a very basic starting point for creating an accurate Building Information Model with which to continue planning. Despite this limited overall completion, it should be recognised that this is groundbreaking technology in terms of its complexity as a piece of software. The ability to recognise and classify point cloud data into building components is in relatively recent advancement and is primarily used in the MEP (mechanical electrical and plumbing) discipline.
Being given accurate external walls as a starting point for as built model creation, is a definite benefit to planning. The ability to produce ‘usable’ native Revit system families is ahead of the curve in terms of what much state-of-the-art post processing of point cloud data software can achieve at present.
The manual process was found to be time consuming in terms of being completely un-automated in geometry generation, however, based on the competence of the user, it is possible to get a highly accurate 3D model using the point cloud as an underlay. This functionality in Revit has come a long way in the past few years and makes importing scan data and working natively in Revit, very user friendly.
Which process to use? In conclusion, it depends! There is still a long way to go before we will have fully automated BIM creation from scan data. The best case at the moment seems to be using a mix of both methods, taking automated geometry where possible and manually altering it for the purposes of the specific project.
Point Cloud Received
Last week I received the registered point cloud from the very brilliant Farrimond MacManus LTD containing literally millions of digital points, each on an X Y Z axis in space. Point clouds are the outcome of a 3D laser scan and are used in the construction industry for planning construction and maintenance work with absolute precision.
Designers can use the point cloud in place of an external reference, or in other words, an underlay in their design software showing the layout of the existing conditions of their site. By slicing up the cloud, sections of the building elements and terrain can be examined in detail.
It was amazing to find issues with the building that would have been very difficult to see without this technology. For example, we found a section of the wall, at the meeting of the main cottage with the back room, to have a joining wall of just a single layer of stones (just 150mm compared to the usual aprox. 600mm elsewhere!). These important details will be very significant in planning for moisture and damp aversion.
The point cloud can also be used in place of return visits to site. Navigating through the points in Autodesk ReCap or other software, measurements or details that may have been missed on survey day can be checked, again, with astonishing accuracy. By mapping photographs, taken during the scanning process, onto the cloud we can get a photo realistic walk through and exploration of the buildings and surrounding area.
The next step in the process will be to translate this cloud of unintelligent points, into a building information model (link), upon which semantic data can be associated to facilitate construction design and documentation. For the research I will be looking at the state of the art in automating this process using component recognition software as well as producing the BIM manually as is the usual practice in the industry at present.