Clash Detection in BIM Modeling

February 28, 2012 — 18,485 views  

Clash detection is an important and integral part of the BIM modeling process. Clash detection arises out of the fact that, in BIM modeling, there is not just one model, but several, that are, in the end integrated into a composite master model. Each discipline: structural engineering, MEP engineering, environmental engineering, etc…, creates a model, independently of all the others, based upon the architects original model, which is the starting point for all the other disciplines. After each of the disciplines has finished their work, the next step in BIM modeling is clash detection, which is the process of finding where the models "clash": elements of separate models occupying the same space, or with parameters that are incompatible, or in 4D BIM modeling, a time sequence that is out of order. Finding these inconsistencies is vital, as they would severely impact the construction process, causing delays, design changes, materials costs and a cascade of headaches and budget overruns.

Clash detection is not new; it's just that, in the past, clash detection took place on the construction site, when the beam that the structural engineer designed is right in the path of the air conditioning units the MEP engineer located. Huge expenses and costly delays were necessary to fix this "clash detection". In BIM modeling, clash detection takes place during the design phase, so that constructability issues can be resolved before construction begins, saving vast sums of money, time and producing a better building. Prior to computerization, clash detection was a manual process of overlaying drawings on a light table to visually see clashes. 2D CAD operated essentially the same way, which did nothing to help in this arena. It was 3D graphics that allowed for the inherent detection of clashes, though in the earliest stages, this was often still done visually by the engineer, or if automated returned many meaningless clashes, and since early on these objects were really only defined surfaces, detection of objects within objects was impossible. So early 3D was still time consuming and not even close to 100% accurate.

The advent of true BIM modeling included the integration of clash detection search capabilities in the software, which tremendously enhanced the process, both in terms of speed and accuracy. These methods actually have been borrowed from the gaming community, where "collision detection" has been a feature of virtual reality for a long time. Can't have monsters running through each other! But the real-time character of games places their emphasis on speed, at the cost of accuracy. In BIM modeling, it's accuracy that counts, though speed is important, but distinctly secondary.

The example below gives a visual, 2D idea of how BIM modeling clash detection algorithms work. Fundamentally, a simpler shape is drawn around each object and then the program checks to see if there is a geometric/spatial overlap. The simpler the shape, the faster the BIM modeling algorithm can analyze the model

Courtesy TNO Built Environment and Geosc 1

Comparing spheres (A & B above) is the fastest algorithm. Two spheres collide if the distance between the centers of the two spheres is smaller than the sum of the radiuses of the two spheres. Bounding Box (C & D above) is more accurate but slower. Clash detection of the actual BIM modeling shapes ( E & F above) uses an intensely more difficult algorithm, but correspondingly more accurate.

Most BIM modeling clash detection software uses a combination of these, and other (triangle/ray) methods, by defining model objects into classes based upon the likelihood of clash. Elements that have a low likelihood of clashing use the faster methods, while objects, such as floor and ceiling beams, use the more accurate geometries, optimizing the process for both speed and accuracy.

Clash detection in BIM modeling looks for three classes of clashes:

Hard Clash

Soft Clash\Clearance Clash

4D\Workflow Clash

Hard clash is exactly what it seems; two objects occupying the same space: a beam where a plumbing run is designed, a column running right through a wall. Simple stuff. But BIM modeling hard clash detection brings not only geometry based detection, but semantic and rule-based detection algorithms, due to the embedded information in the BIM modeling objects. Geometry based clash detection will return a clash for every recessed ceiling light, or a pipe running through a wall. But clash detection rules based on embedded object data can eliminate these common mistakes. The level of detail in BIM modeling is extremely important for the accuracy and efficacy of clash detection. So are selection sets, which allow a BIM modeling user to run a clash detection between specific subsets of a model, such as MEP against just ceilings, or structural against just walls.

Soft Clash/Clearance refers to objects that demand certain spatial/geometric tolerances or buffers having objects within their buffer zone for access, insulation, maintenance or safety. Soft clashes are one of the real avenues where BIM modeling clash detection has brought new capability to the process. Custom "soft clash" detection can go as far as checking components for building code adherence, based, of course, on a robust object data population.

4D/Workflow clash detection refers to the ability of a BIM modeling project to resolve scheduling clashes for work crews, equipment/materials fabrication and delivery clashes and other project timeline issues.

After running a clash detection scan, there are usually many duplicate instances of the same clash. A plumbing line running through 8 wide-flange beams will show up as 8 clashes, though in reality it's a single issue. Condensing and culling clashes is a normal part of the BIM modeling clash detection process.

Clash detection technology breaks down into two arenas: (1) clash detection within the BIM modeling design software and (2) separate BIM integration tools that perform clash detection. Clash detection in instance (1) is limited due to the limited ability of most BIM modeling design software to integrate multiple, non-proprietary models. The norm, however, is that different disciplines of the design and construction team will do their work on different software platforms. The structural team may be using Tekla, the architectural model may be built using Revit, the electrical contractor may use Bentley, and the HVAC engineers may deploy Graphisoft. These applications do not speak directly to each other and so cannot alert one another that clashes occur. So instance (2) makes the most sense and also generally has more powerful and sophisticated clash detection tools. However, the drawback of most BIM integration software is that clashes cannot be fixed within the integrated model. Interoperability issues render this software mostly one-way. Models can be imported, but not altered and exported. As the non-proprietary, open-specification interoperability IFC model specification gains more credence, this issue will be resolved.

The importance of clash detection in BIM modeling is hard to overstate. It has been estimated that, industry-wide, each identified clash saves about $17,000 on a project. On large projects, 2000-3000 clashes are not unusual. That's $34,000,000! So understanding and using clash detection elements in BIM modeling software is crucial for engineers, architects, owners and contractors.



O'Donnell & Naccarato