4.2.1. Initial Situation
The first variation considers that no information is given regarding the interdependencies between elements. Therefore, the DNA method, in accordance with the guiding rules, considers the worst case, i.e., all elements depend on the structural elements (elements e9 and e10) (Figure 5
If the element to recover is the left beam (e7), a quick preliminary check gives the following results:
Therefore, the preliminary check concludes that e7 is potentially recoverable, and the assessment can continue. The next step establishes the potential paths. Because many elements are accessible, numerous paths can be defined (Figure 6
), such as:
Path P1: elements e7, e10 and connection c9
Path P2: elements e7, e4 and connection c2
Path P3: elements e7, e5 and connection c3
Other paths contain one of the three previous paths and are longer.
In path P1, several elements must be removed or connection must be unfastened due to interdependencies (Figure 7
However, because these elements are also hosting others, the list of elements and connections grows until all elements and connections are considered:
elements e10, e4, e1, e4, e2, e5, e3, e6 e8, e9, e7
connections c9, c1, c2, c7, c4, c6, c5, c3, c8
Any other path will be equivalent to path P1 because of the lack of information concerning the interdependencies between different elements. Here, all connections are considered crucial to the building stability and therefore everything must be dismantled. Consequently, these paths are ignored. This case clearly highlights the need for dependencies as the attempt to disassembly one element of the system induces the removal of every element and connection. Due to limited information, the insight provided is limited to recovered and lost materials and total disassembly time. The answer provided by the tool would only be as follows: “The removal of element e7 (and thus, the removal of everything), generates that amount of waste (total waste to disassemble the whole building) in this amount of time (total disassembly time). Connection c4 induces the loss of element e5 and replacing it by a reversible connection would therefore avoid the loss of e5 and We5.”. No information is provided regarding the disassembly sequence or the impact of removing or repairing one particular object on the whole structure. Additionally, comparison of scenarios and design options are limited to the reduced number of parameters that could be studied. Some lessons learned for the DNA method are as follows:
The final disassembly sequence is not realistic compared to the nature of the elements (beams and columns). In this case, the DNA method always considers that the disassembled objects are supporting the others. In that case, the support (column) is considered as something that needs to be removed prior to the beam (hosted element); this does not come directly from the method but from missing information and data.
In future development, more nuance can be applied to the default value by considering additional metadata stored within BIM objects, such as the type of object (e.g., beam and column) and their location (e.g., level and host).
The distinction between sequential and parallel disassembly is not made and results in longer disassembly time. Indeed, the method calculates the number of working hours needed without considering if workers can work in parallel. Hence, future alternatives to distinguish sequential and parallel disassembly can be considered, although this has not been identified as a necessary feature.
Interdependencies are needed for more detailed investigations and results.
4.2.2. Path Determination on a Simplified Case Considering Dependencies
The second variation considers that nonstructural elements depend on structural ones. If two structural elements are connected, the designer must define if they are interdependent. In this case, the beam e7 depends on columns e4 and e5. However, columns e4 or e5 do not depend on e7. Ducts e9 and ceiling e10 depend on e7. Considering the same problem as in the first variation, only two potential sequences to recover e7 are considered (Figure 8
In path P1: elements e7, e10 and connection c9
In path P2: elements e7, e4 and connection c2
Considering the interdependencies, the final list of elements and connections for each path are:
In path P1: elements e10, e9, e7 and connections c9, c8, c3, c2 (Figure 9
In path P2: elements e4, e10, e9, e7 and connections c9, c8, c3, c2, c1
As a first remark, if ducts e10 was inaccessible, the disassembly of e7 would be impossible because e9 and e10 should be removed prior to recovering e7. This would mean that the path considered is not viable, and the method will delete this path from the potential ones.
The user knows the potential ways to recover e7 in addition to a simplified disassembly sequence that could be used to phase the disassembly. If continued, the analysis would allow the time and amount of recovered and lost materials per assembly to be quantified. Below are some lessons learned for the DNA method:
Following the second guiding rule, this case considers the (structural) interdependence between elements, which therefore reduces the amounts of elements to be removed (compared to case 1). By adding extra information, the user is rewarded with a reduction of the disassembly time and waste generated.
Two different paths show different results, which illustrates that the method can consider and distinguish various paths. This also represents a huge opportunity because, while a manual method might be more efficient and effective for a reduced number of paths, the added value of the DNA tool will be in the evaluation of several paths and disassembly sequences as a whole.
No distinction is made between e9 and e10 although, in reality, e10 must be removed prior to e9. This shows the necessity to add a factor considering the “dependence regarding the accessibility of the element”. Indeed, connection c8 and element e9 are only accessible after the removal of e10. This can be made by adding accessibility interdependencies (similar way to structural dependency) as inputs, either as a parameter in the model, a cell in a calculation sheet, or through a dependency graph.
4.2.4. Complete Application of the Method on the Frame Structure
Supposing the aim is to get element e2, which is transportable, we first need to check if it is possible to extract e2 before checking how we can access it. The connection c4 is not reversible; thus, the method must establish if e2 could be damaged by the removal of this connection. To do so, the resistance (theoretical value, which depends on several factors that should be further investigated) between the element e2, the connection, and the other element it is connected to are compared. In this case R2, Rc4, and R5 are compared with the resistance of the elements and the resistance of the in-between connection. Supposing the resistance of e2 (R2) is higher than at least one of the others, e2 can be removed without damage.
Element e2 is considered directly accessible (blue in the network), which suggests that it will be easy to get it. If this path exists, it will always be prioritized because it only considers the removal of connections without having to remove extra elements. Therefore, all other paths will at least include this subpart. Potential paths are as follows:
However, due to the structural interdependencies elements, e10, e9, e7, e8, and e5 must be removed prior to e2. And thus, the disassembly sequence (by connection) is c9, c8, c2, c3, c5, c6, c4 (Figure 11
As connections c9, c8, c3, c2 are reversible (and in the right direction), elements e10, e9, ande7 are added to the recovered elements list. The time considered to remove these connections is the disassembly time (Td), which is 60 min, 30 min, 45 min, and 45 min for c9, c8, c3, and c2, respectively.
On the contrary, all the other connections are not reversible, which means that waste will be created. However, to know which elements are going to be demolished, we need to compare their resistance.
Considering that Re2 > Re5 = Re6 > Re8 and that the connection itself will never break(if the connection breaks then no waste is generated except the connection itself), it induces that the removal of c5 will break e8, which will be included in the list of lost elements with a destruction time Tb = 10 min. Then, the removal of c6 is the same. The connection c4 is then removed, which generates the loss of e5 (in 5 min).
In the end, the lost elements are e5 and e8, and the total weight of waste is the sum of each element’s weight. The total disassembly time is the sum of the disassembly times (Td) except for c4, c5, and c6, where the destruction time (Tb) has been considered.
For path P1: Total disassembly time = 60 + 30 + 45 + 45 + 10 + 10 + 5 = 205 minutes.
Total amount of waste = We5 + We8.
Additionally, the method informs the user that the loss of e5 is only due to connection c4. Therefore, if c4 is replaced by a reversible connection, the total amount of waste will be reduced by We5. However, this does not apply for element e8 as it is broken due to connections c5 and c6.
As an alternative result, connection c9 takes 60 min (Td) to be unfastened because the ceiling is attached in several places to the beam, but it takes only 10 min to break (Tb). Therefore, if time is a predominant factor, the method might suggest the following: “In path P1, if you remove c9 by breaking e10, the disassembly duration will be 50 min shorter (25%) but you will generate 25 kg of extra waste (We10)”.
Some lessons learned for the DNA method:
The removal of connections c5 and c6 both generate the loss of element e8. Element e8 is listed twice in the lost elements list to differentiate this case with a case where only one connection is leading to the loss of e8. However, when calculating the amount of waste generated by one element, the weight will be counted only once. The time of each disconnection is considered as we do not know whether the two connections will fail together or not (to be on the safe side).
While checking connection c4, it is important to note that although the target element is the weakest, it may still be possible to get a part of it. In the method, the elements are considered “recovered”’, only if the conditions for disassembly are met. Therefore, we have decided to overestimate waste and underestimate elements that could be recovered (to be on the safe side).