Another interesting perspective in modern museum work is connected with the use of 3D scanning for the reconstruction of damaged and even completely destroyed monuments.
3.1. Case Study of Zinc Sculpture “Eve”
One of the case studies of such works is the reconstruction of a totally lost sculpture from the former estate Sergievka in the suburbs of St. Petersburg, which in the XIX century was the country residence of the Duke of Leuchtenberg, the stepson of Napoleon Bonaparte. The Duke of Leichtenberg collected antiquities and possessed the richest collection of masterpieces. Unfortunately, almost all the items of this unique collection were lost during the Great Fatherland War, as Sergievka was in the epicenter of combat activities. Among the many lost masterpieces was a 19th-century zinc sculpture “Eve at the water spring”, the fragments of which were discovered underground during construction works in 2007 (
Figure 2).
They represented a pile of scattered fragments that did not allow carrying out the reconstruction of this monument by traditional methods of the museum work. However, the use of 3D laser scanning technology, computer simulation, and 3D printing made possible creating a replica of this sculpture, allowing us to judge what the original monument was like (
Figure 3). The replica was created from polymer material using additive technology FDM (fused deposition modeling).
The results of this project are described in the work [
5], but it is worth mentioning that the replication and reconstruction of works of art puts a serious issue on the agenda—what is the accuracy of creating the physical copies of objects? In other words, how precise is the correspondence between surfaces of the same object in its 3D model and replica obtained using additive technologies? This issue is very important from point of view of authenticity of any original CH object and its replica, but it is not considered in the scientific literature in the papers devoted to the replication of sculptures.
We carried out an experimental study on the evaluation of the accuracy of the replication of the sculpture “Eve”. In those experiments, we used a laser scanner Surphaser 25SHX (scanning accuracy—2.5 microns) and did find that deviations in the surfaces of the original 3D model and replica were in the range of ±0.7 mm. During a detailed analysis of the 3D model of the replica, a “cellular structure” (resembling honeycomb) was clearly visible on its surface, especially pronounced in the area of the sculpture’s back (see
Figure 4, where one can see the result of “overlapping” of 3D model of replica and 3D model that was used for the creation of replica).
This may be a consequence of the 3D printing technique used, in which, to reduce the cost of replication, we decided to save used material (PLA plastic) by having the sculpture be hollow inside (the volume filling was 10%). Because of this, local “minima” (deflections) of the sculpture surface are observed in the places where the surface layer of the sculpture is attached to the stiffeners, while the sagging areas correspond to local maxima. It was these deformations that were detected by laser scanning.
Thus, the accuracy of reproduction of the sculpture “Eve” can be considered good, but it must be borne in mind that when replicating sculptural monuments, distortions of their geometric shape are possible.
3.2. Case Study of Cast-Iron Star
One of our most recent works relates to the reconstruction of CH objects created from cast-iron. Nowadays, it is a typical conservation problem in CH preservation. There are sculptures, objects of decorative art, grids around historical buildings, parks, and gardens as well as just decorative elements on the facades of buildings. Many of them are highly deteriorated and usually have severely damaged and lost parts (see
Figure 5a). Until recently, experts in the field of CH did not attach much importance to such items, and as they fell into complete disrepair, they were replaced with copies. However, in recent years, there has been a reassessment of values, and a new trend has emerged in the work of restorers, which requires the reconstruction of objects damaged by corrosion instead of replacing them with copies.
The authors of this article understand this problem and proposed its technical solution based on the use of laser powder coating technology. In our work, we demonstrated the principal possibility of the realization of such an approach.
The object of our studies was a small (10 cm) iron star that is a decorative element of a XIX century cast-iron tombstone fence of the Alexander Nevsky monastery in St. Petersburg. This object is highly deteriorated due to corrosion and has losses of some small elements (see
Figure 5b).
The main idea of our project was to recreate the losses of individual elements of the star using a combination of 3D scanning technology, direct metal laser sintering (DMLS), and laser cladding.
Firstly, we analyzed the chemical composition of the star using X-ray fluorescence using XL3t-32280 equipment. The analysis results are shown in
Table 1.
We did find that the main chemical element is Fe (its concentration is of about 91%). Additionally, there are permanent impurities typical for gray cast-iron (the total concentration is only about 7.2%).
Secondly, we removed the corrosion from the star surface. For the removal of corroded layers, we tried to use different treatment techniques: laser cleaning, sand blasting and chemical treatment; the latter gave the best result (see
Figure 6a).
Then, we carried out 3D scanning of the star using laser triangular scanner Konica Minolta V-910 and created its 3D model. The next stage of our work was the computer modeling. The missing ray of the cast-iron star was modeled with the Zbrush software. The main distinguishing feature of this software is the ability to “sculpt” 3D objects. The surviving cones were projected, processed, and “glued” to the main body of the star in the parts where there were losses (see
Figure 6b), where blue ends are reconstructed lost parts of the star.
It was decided to separate those experiments into two stages: firstly, we created one of the star rays, which we proposed to join with the star; the second stage of reconstruction will be based on a reconstruction of the lost end of another star ray by direct metal deposition. It is obvious that such a method of reconstruction is more complicated since the very precise movement of a laser beam is needed, and such a task can be solved by means of the development of specialized software.
A separate ray of the star was grown by the DMLS method according to the CAD model. We used a CW fiber Ytterbium laser with an output power of up to 80 W. Stainless-steel powder 316 L20 (manufacturer—Hoganas, Belgium) with a granule size of 53 µm was used for printing. The result of creating this piece can be seen in
Figure 7a.
We then applied direct laser deposition (DLD) [
9,
10] using the OKTA-Printer machine developed at the Peter the Great St. Petersburg Polytechnic University (see
Figure 8). The essence of the DLD method is to directly feed a metal powder or composite powder mixture into a molten bath formed by a laser beam [
11,
12,
13]. The local action of heating the laser beam, the selective effect, as well as the high accuracy of modern robotic manipulators allow using this technology as an advanced and modern method for repairing and restoring products. The main parts of the OKTA-Printer are the powder cladding head and the CAMAU robot. For joining, we used a nickel alloy powder Inconel 625 (manufacturer—Hoganas, Belgium) with a particle size distribution of 50–150 µm. It is one of the most popular alloys in laser cladding and DLD [
14,
15]. To connect the two pieces, they were glued to each other with glue, then welded in one pass on each side. The following parameters were used for welding: laser spot width 1.5 mm, laser power 700 W, robot movement speed 1500 mm/s, and powder feed 20 g/min. The result of the stage of the star reconstruction is shown in
Figure 7b.
Using the DLD printer, an attempt was made to deposit the missing part of the star’s ray on the material from which it is made. For this, a parallel piece was cut off and deposited on it from the same Inconel 625 that was used for welding. As a result of the cladding, an excessive absorption of laser radiation by the star material was detected, which leads to melting and a geometry change in the substrate along the edges, even with a relatively low heat input.
Furthermore, to assess the thermal effect on the structure of the star material, a materials science study of the initial structure and structure after laser cladding was carried out, and the transition zone was also investigated. Additionally, the microhardness was measured at a load of 0.3 kgf with the same step of 200 µm to evaluate the change in metal properties from the nickel alloy overlay to the cast-iron from which the star is made (see
Figure 9).
Separate hardness measurements were made at four to five points for each area of interest: cast-iron without surfacing is 240 ± 60 HV, with weld overlay 467 ± 76 HV, hardness of deposited Inconel 625 corresponds to 462 ± 26 HV, hardness in the transition zone is 601 ± 134 HV. Examination of the state of the interface on an optical microscope (
Figure 9c) showed a strong metallic bond of the cladding with the star material: there were no cracks or pores at the interface.
Since there are no data on the percentage of carbon in the composition of the studied cast-iron, we assume that the initial microstructure of the star material corresponds to the structure of gray cast-iron on a ferrite base with lamellar-free graphite. Due to the thermal effect of the laser beam, the hardness of cast-iron increased significantly, which can be associated with rapid heating followed by rapid cooling, which leads to the formation of a pearlite structure, as well as the accumulation of thermal stresses. It was noted that at more than 1 mm from the surfacing, the hardness of the cast-iron base was significantly higher than the hardness of the same cast-iron without cladding. This indicates a large heat-affected zone that is not typical for standard cladding materials. Usually, this area is no more than 0.5 mm. This is most likely due to the size of the substrate itself and the low heat-dissipation. For the experiment on direct restoration of the missing part, a substrate with an area of 2 cm2 was used, whereas the substrates for cladding and DLD are usually at least 100 cm2.